青藏高原植被返青和枯落期（2015-2100）预测数据集

植被物候对气候变化响应敏感。为探索未来气候变化情境下，青藏高原植被返青期和枯落期的变化轨迹，本数据集基于返青期和枯落期物候模型，模拟预测了CMIP6 15种气候模式、4种二氧化碳排放情境下青藏高原像元尺度2015-2100年的返青期和枯落期。其中，返青期模拟基于构建的长短时记忆（LSTM）深度学习模型进行预测，该模型已被证明能够较高精度地模拟返青期年际变化（R=0.92，详见引用文献）；枯落期模拟基于基于过程的物理模型，该模型在多模型比较中具有模拟青藏高原枯落期的精度优势（详见引用文献）。本数据能够用于模拟未来生态系统变化的物候输入参数，对青藏高原生态系统的保护、植被物候对于气候变暖的响应以及指导农事活动等方面具有一定的参考意义。

Vegetation phenology is highly sensitive to climate change. To explore the temporal trajectories of the start of growing season (SOS) and end of growing season (EOS) of vegetation on the Qinghai-Tibet Plateau under future climate change scenarios, this dataset simulates and predicts the SOS and EOS at the pixel scale of the Qinghai-Tibet Plateau from 2015 to 2100, utilizing 15 CMIP6 climate models and 4 CO₂ emission scenarios based on phenological models for SOS and EOS. Specifically, the SOS simulation is performed using a developed long short-term memory (LSTM) deep learning model, which has been validated to accurately reproduce the interannual variations of SOS with a correlation coefficient R=0.92 (see cited references for details). The EOS simulation is based on a process-based physical model, which demonstrates superior accuracy in simulating the EOS of the Qinghai-Tibet Plateau in multi-model comparisons (see cited references for details). This dataset can serve as phenological input parameters for simulating future ecosystem changes, and has certain reference significance for the conservation of the Qinghai-Tibet Plateau ecosystem, research on the response of vegetation phenology to climate warming, and guiding agricultural production and other related fields.